Statistical/practical issues in clinical trials

For approval of a drug product, the United States Food and Drug Administration (FDA) requires that substantial evidence of the effectiveness and safety of the drug product be provided through the conduct of two adequate, well-controlled clinical trials. To assist the sponsors in preparation of final clinical reports for regulatory submission and review, the FDA and other regulatory agencies and conferences such as the International Conference on Harmonization (ICH) have developed guidelines for the format and content of a clinical report. The FDA and ICH guidelines require that the following statistical issues be addressed in the final clinical report: 1. Baseline comparability, 2. Adjustments for covariates, 3. Dropouts or missing values, 4. Interim analyses and data monitoring, 5. Multicenter studies, 6. Multiplicity, 7. Intention-to-treat subset versus efficacy subset, 8. Active control trials, and 9. Subgroup analyses. This paper provides an overview of these statistical issues. In addition, statistical justification for these issues is also addressed. © 1997, Drug Information Association. All rights reserved.

Duke Scholars

Author Shein-Chung Chow Biostatistics & Bioinformatics, Division of Biostatistics